Common teaching processes like the teach pendant control most industrial robots. Accelerometer-based gesture recognition has become popular since the last ten years. Advantages of the accelerometer are low to moderate in cost, and small in size. Wireless sensor system (using 'Wii' remote; accelerometer-based) can be used to control a robotic arm. This arm is designed to operate in similar motions to the human arm.
Calibration of these teaching methods require large amount of time. They can require extensive human intervention. Adjustments are done based on accelerometer data. When teaching a walking robot (bipedal), feedback is useful to learn new walk parameters.
3-axis accelerometer is attached to various parts of a robotic arm. Sensor (accelerometer) is configured to recognize gestures (changing arm positions). Only specific movements can activate the robotic arm to function. Accelerometer data is analyzed to fit in the recognized specific movements, eg, a swing to the right. Slight jittery movements will not produce any meaningfulful specific patterns. Advantage of such control system is its programmable, repeatable robotic arm movements.
Two 'Wii Nunchucks' (3-axis accelerometer) can be attached to control a robotic arm. One is at the palm of the user (held), another at his elbow. By this, at the two end points of the arm (palm and elbow), differentiation between roll and pitch wrist and shoulder joint movements are allowed. Digital motion data is then sent to (from both 'Wii Nunchucks') to 'Arduino' micro controller. The 'Arduino' micro controller only collects and processes acceleration data. A separate micro controller then controls the servo motors. 'Arduino', mentioned many times in this article is a physical computing platform that is currently taking the world by storm.
Source by Alfred Chai Wei Liang